Va Decision Support Seminars 2009

Decision Support in VA Decision Support in VA Clinical Information Clinical Information

SystemsSystems

James L. Breeling, MDJames L. Breeling, MD

March 18, 2009March 18, 2009

BU Informatics SeriesBU Informatics Series

Diffusion of InnovationsDiffusion of Innovations

Classical diffusion theory assumes that Classical diffusion theory assumes that innovations are sufficiently well defined for innovations are sufficiently well defined for individuals to assess their relative individuals to assess their relative advantage, risk and cost. advantage, risk and cost.

In reality, innovations carry uncertainty for In reality, innovations carry uncertainty for individuals and therefore delay in adoptionindividuals and therefore delay in adoption

Adoption of medical innovation in Adoption of medical innovation in particular is slower and less predictableparticular is slower and less predictable

Clinical Knowledge is a Special Clinical Knowledge is a Special CaseCase

Physician expertise is built through Physician expertise is built through Acquiring specific scientific knowledgeAcquiring specific scientific knowledge Learning-by-doingLearning-by-doing

Practice Communities tend to be local Practice Communities tend to be local “closed” systems despite links to “closed” systems despite links to international knowledgeinternational knowledge

General trends in local communities are General trends in local communities are shaped by opinion leadersshaped by opinion leaders

Unexplained Practice VariationUnexplained Practice Variation

Business Drivers for Evidence Business Drivers for Evidence Based MedicineBased Medicine

Employers and consumers are now paying Employers and consumers are now paying for performancefor performance

Institute of Medicine report on medical Institute of Medicine report on medical errorserrors

Prescribing datasets suggest that 5-10% Prescribing datasets suggest that 5-10% of medication prescriptions lead to of medication prescriptions lead to preventable, anticipatable adverse eventspreventable, anticipatable adverse events

Guideline ImplementationGuideline ImplementationPractices or Practices or StrategiesStrategies

DefinitionDefinition

Organizational & Organizational & AdministrativeAdministrative

Changes in Leadership, responsibility, policy, Changes in Leadership, responsibility, policy, decision making, culture, knowledge management, decision making, culture, knowledge management, communication and coordinationcommunication and coordination

Technology & Technology & InfrastructureInfrastructure

Changes in information technology, data Changes in information technology, data management, feedback, physical plant and management, feedback, physical plant and equipmentequipment

Care DeliveryCare Delivery Changes in care delivery process, patient Changes in care delivery process, patient interactions, patient education, clinical programs and interactions, patient education, clinical programs and treatment analysistreatment analysis

Staff Resources & Staff Resources & DevelopmentDevelopment

Changes in resource allocation, increased awareness Changes in resource allocation, increased awareness and knowledge, skill building, influence provider and knowledge, skill building, influence provider attitude, behavior changes.attitude, behavior changes.

VHA ORD HSR&D Blueprint for Change Jan 26, 2000

Expansion of Medical KnowledgeExpansion of Medical Knowledge

““The surge of new knowledge in medicine The surge of new knowledge in medicine has created a tremendous problem for the has created a tremendous problem for the student and for the practitioner. The student and for the practitioner. The periodical literature, for a long time periodical literature, for a long time unwieldy, has now become impossible to unwieldy, has now become impossible to handle…”handle…” Arthur L. Bloomfield, MD, “A Bibliography of Arthur L. Bloomfield, MD, “A Bibliography of

Internal Medicine” 1957Internal Medicine” 1957

Clinical Decision Support Clinical Decision Support ChallengesChallenges

Meaningful decision support should Meaningful decision support should improve quality of care for patientimprove quality of care for patient

Meaningful decision support should Meaningful decision support should improved quality of clinician interaction improved quality of clinician interaction with electronic recordwith electronic record

Meaningful decision support will be based Meaningful decision support will be based on highest quality knowledge base in a on highest quality knowledge base in a timely fashion (“don’t be the first or the last timely fashion (“don’t be the first or the last to prescribe a new drug”)to prescribe a new drug”)

Facets of Clinical Decision Support Facets of Clinical Decision Support Quality & Performance

•Evidence based•Safety based•Regulatory based

Patient Perspective•Preferred agent•Preferred dosing•preferred route

End User Experience•End user role•Office workflow•Preferred workflow

Pitfalls of Clinical Decision SupportPitfalls of Clinical Decision Support

““task-interference” as a method of guideline task-interference” as a method of guideline enforcementenforcement Make it difficult for clinician to complete a process Make it difficult for clinician to complete a process

unless minimum threshold metunless minimum threshold met Decision Support through interference with taskDecision Support through interference with task

““task-facilitation” as a method of guideline task-facilitation” as a method of guideline enforcementenforcement Manage disease states through presentation of Manage disease states through presentation of

context sensitive rules, order sets, documentation context sensitive rules, order sets, documentation templatestemplates

Decision support should facilitate goal – i.e. manage Decision support should facilitate goal – i.e. manage Diabetes for better glycemic controlDiabetes for better glycemic control

Knowledge Life-CycleKnowledge Life-Cycle

Proposed Guideline & Content

Guideline Defined & Validated

Functional Knowledge Specified for Encoding

Engineered into Production

Ongoing Revisions&

Eventual Sunset

VISTA Clinical Reminder Package: VISTA Clinical Reminder Package: Designing a ReminderDesigning a Reminder

Think of the clinical practice guideline as Think of the clinical practice guideline as an algorithman algorithm

Think of the clinical reminder as a decision Think of the clinical reminder as a decision node in the algorithmnode in the algorithm

State in plain English the target State in plain English the target population=“denominator”population=“denominator”

State in plain English the action that State in plain English the action that resolves the reminder=“numerator”resolves the reminder=“numerator”

VISTA Clinical Reminders: FindingsVISTA Clinical Reminders: Findings

DrugDrug Education topicEducation topic ExamExam Health factorHealth factor ImmunizationImmunization Laboratory testLaboratory test Mental health instrumentMental health instrument Orderable itemOrderable item Radiology procedureRadiology procedure

Reminder computed Reminder computed findingfinding

Reminder taxonomyReminder taxonomy Reminder termReminder term Skin testSkin test VA drug classVA drug class VA genericVA generic Vital measurementVital measurement Reminder location listReminder location list

Boolean Logic: How Findings Boolean Logic: How Findings Are Used to Define PopulationsAre Used to Define Populations

A AND NOT B

A B

A AND B

A B

A OR B

A B

The negation of “A or B” is “Not A and not B”

The negation of “A and B” is “Not A or not B”

Importance of DenominatorsImportance of Denominators

A finding that make a reminder not A finding that make a reminder not applicable for a patient, causes that applicable for a patient, causes that patient to NOT be included in the patient to NOT be included in the denominatordenominator patient is not in the cohortpatient is not in the cohort

A finding that satisfies or resolves the A finding that satisfies or resolves the reminder leaves the patient in the reminder leaves the patient in the denominator.denominator.

Two Difference Populations Two Difference Populations Receiving PneumovaxReceiving Pneumovax

1 Age AND Sex OR High RiskAge AND Sex OR High Risk satisfied by vaccine or refusalsatisfied by vaccine or refusal

2 Age AND Sex OR High Risk Age AND Sex OR High Risk AND NOT Refusal AND NOT Refusal

satisfied by vaccinesatisfied by vaccine

Pneumovax Example:Pneumovax Example:DenominatorsDenominators

# Patients with# Patients with Reminder Applicable Reminder Due Reminder Applicable Reminder Due ------------------- ------------ ------------------- ------------

1 Pneumovax 33 111 Pneumovax 33 11

# Patients with# Patients with Reminder Applicable Reminder Due Reminder Applicable Reminder Due ------------------- ------------ ------------------- ------------

2 Pneumovax 19 112 Pneumovax 19 11

45 patients evaluated45 patients evaluated

Dialogs: What Is in the Dialogs: What Is in the Programming Tool Box?Programming Tool Box?

Radio buttonsRadio buttons one choice only from group of responsesone choice only from group of responses

Check boxesCheck boxes multiple choicesmultiple choices

Hidden/visible itemsHidden/visible items Informational textInformational text

no user entryno user entry Free text comment boxesFree text comment boxes

user entry optional or requireduser entry optional or required Forced YES/NO answersForced YES/NO answers HyperlinksHyperlinks

Dialogs – Informatics Dialogs – Informatics ErgonomicsErgonomics

Bad practicesBad practices Too many mouse Too many mouse

clicks per window to clicks per window to process dialogprocess dialog

Window scrolls down Window scrolls down in seemingly unlimited in seemingly unlimited fashionfashion

Too many embedded Too many embedded windows appear in windows appear in successionsuccession

Good practicesGood practices Present just enough Present just enough

useful options to user useful options to user to processto process

Allows user to quickly Allows user to quickly process relevant process relevant information and then information and then return main levelreturn main level

ApplicationsApplications

Performance improvementPerformance improvement Decision SupportDecision Support Case FindingCase Finding Risk ProfileRisk Profile

PACTPACT MOVEMOVE

Standards of CareStandards of Care

Patient ListsPatient Lists

New in Clinical Reminders V.2.0New in Clinical Reminders V.2.0 Created as part of reminder extractsCreated as part of reminder extracts

Patient lists can be built independently of extractsPatient lists can be built independently of extracts

Uses Clinical Reminders IndexUses Clinical Reminders Index Fast compilingFast compiling

Uses?Uses? Extracts, Health Summaries, Reminder Reports, Extracts, Health Summaries, Reminder Reports,

Demographics, Patient Mailing ListsDemographics, Patient Mailing Lists

Clinical Reminders – Network Clinical Reminders – Network Initiative and GovernanceInitiative and Governance

Oversight of Clinical Reminders is by the Clinical Practice Oversight of Clinical Reminders is by the Clinical Practice Guidelines committeeGuidelines committee

Every reminder must have a service line sponsor Every reminder must have a service line sponsor (administrative or clinical)(administrative or clinical)

Every reminder must have an owner who is responsible for Every reminder must have an owner who is responsible for the content and for awareness of need for changethe content and for awareness of need for change

A subcommittee accepts proposals for new reminders, A subcommittee accepts proposals for new reminders, recommends priorities, may help develop prototypes and recommends priorities, may help develop prototypes and conduct alpha testsconduct alpha tests

An implementation group meets quarterly to roll out An implementation group meets quarterly to roll out reminders or modifications on a uniform basisreminders or modifications on a uniform basis

Clinical Decision SupportClinical Decision Support

Point-of-care decision support Point-of-care decision support Provides information to the clinician at the Provides information to the clinician at the

time of medical-decision-makingtime of medical-decision-making More than clinical remindersMore than clinical reminders

More complex reasoningMore complex reasoning Evidence-base for recommendationsEvidence-base for recommendations

Clinical Decision SupportClinical Decision Support

Active systemsActive systems Drug allergy alertsDrug allergy alerts Lab alertsLab alerts Abnormal x-ray Abnormal x-ray

findingsfindings

Passive SystemsPassive Systems Problem listsProblem lists Intelligent order sets Intelligent order sets

based on evidencebased on evidence Antibiotics listed for Antibiotics listed for

specific conditionsspecific conditions Social history specifics Social history specifics

triggers social work triggers social work consultconsult

Family history Family history suggests specific suggests specific testing or screeningtesting or screening

Principle 1: If the Information Is Principle 1: If the Information Is Necessary, It Should Be Made Necessary, It Should Be Made

Known Before Any ActionKnown Before Any Action Allergies should be listed in all med Allergies should be listed in all med

order screensorder screens Renal function should be visible when Renal function should be visible when

ordering nephrotoxic drugsordering nephrotoxic drugs formulary drugs for a given condition formulary drugs for a given condition

should be listed firstshould be listed first

Principle 2: Intrusiveness is Principle 2: Intrusiveness is more tolerable when the more tolerable when the

situation is more complexsituation is more complex

Knowledge of renal dysfunction as Knowledge of renal dysfunction as related to co-morbidity of CHF or related to co-morbidity of CHF or diabetesdiabetes

Knowledge of HIV treatment regimen Knowledge of HIV treatment regimen and drug-drug interactions and adverse and drug-drug interactions and adverse drug eventsdrug events

Controversies over best diagnostic Controversies over best diagnostic strategies for target conditionstrategies for target condition

The CPR in US Industry: The CPR in US Industry: Gartner Group (1999)Gartner Group (1999)

Fun

ctio

nali

ty

First Generation:The Collector

1993 1998 2003 2008 2012+

Second Generation: The Document

Fourth Generation: The Partner

Fifth Generation: The Mentor

Minimal

Full

Predicted Marketplace Predicted Marketplace MaturityMaturity

Data Acquired Through Feeder Systems

Document Care in Free Text

Structured DocumentationSpanning Episodes of Care

Clinical WorkflowIntervention ScreeningDecision Support

Automated GenerationOf Pathways and Workflow; Use by Consumers

Now

Third Generation: The Helper